create: Create 1- or 2-D non-radial basis

In hillarykoch/nrbfields: Regression-based surface approximation using non-radial basis functions

Description

This collection of functions can be used to generate fields of asymmetric basis functions. Basis functions can be reversed in direction, depending on the data being examined.

Usage

create.weibull.basis(gridlen, shape, scale, nbasis, rev = FALSE, copula = FALSE, copulaType = NULL, param = NULL)
create.gamma.basis(gridlen, shape, rate, nbasis, rev = FALSE,  copula = FALSE, copulaType = NULL, param = NULL))
create.lognormal.basis(gridlen, meanlog, sdlog, nbasis, rev = FALSE,  copula = FALSE, copulaType = NULL, param = NULL))
create.gamma.basis(gridlen, meanlog, sdlog, nbasis, rev = FALSE,  copula = FALSE, copulaType = NULL, param = NULL))
create.weibull.field(gridlen, shape, scale, nbasis, rev=FALSE, reduce = FALSE, copula = FALSE, copulaType = NULL, param = NULL))
create.gamma.field(gridlen, shape, rate, nbasis, rev = FALSE, reduce = FALSE, copula = FALSE, copulaType = NULL, param = NULL))
create.lognormal.field(gridlen, meanlog, sdlog, nbasis, rev = FALSE, reduce = FALSE, copula = FALSE, copulaType = NULL, param = NULL))

Arguments

gridlen	Length of regular grid.
shape	Shape parameter of gamma of weibull distributions.
rate	Rate parameter of gamma distribution.
scale	Scale parameter of weibull distribution.
meanlog	Mean of the normal distribution on the log scale.
sdlog	Standard deviation of the normal distribution on the log scale.
nbasis	Number of basis functions to create in 1 dimension (this gets squared if making a field).
rev	Logical. Should the direction of the basis functions be reverse? In 1-D, this correspond to letting y=f(-x).
reduce	Logical. Should list of 2-D bases be summed into one field?
copula	Logical. Should 2-D bases be made up of dependent 1-D components?
copulaType	A character vector specifying which type of copula should be used to model the dependence when copula = TRUE. Options include "clayton", "frank", "gumbel", "amh", and "joe". These are Archimedean copulas implemented in the copula package.
param	Numeric value of the copula dependence parameter.

Details

Currently, no support for generating irregular grids.

Value

For create.basis when copula=FALSE, a gridlen by nbasis basis matrix is returned. When copula=TRUE, a gridlen by gridlen 2-D basis with only one basis function is returned (this is mostly used for being called by create.field when copula=TRUE). For create field, if reduce = TRUE, a nbasis by nbasis matrix is returned. Otherwise, a list of length nbasis^2, each element containing a gridlen by gridlen matrix corresponding to one basis function, is returned.

Author(s)

[e-mail]

See Also

archmCopula

Examples

twodfield <- create.gamma.field(100, 3, .5, 10, reduce = TRUE)
contour(twodfield)

hillarykoch/nrbfields documentation built on May 25, 2019, 8:17 a.m.